Clinical Landscape of Multi-Specific Antibodies

2:43 

So my name is Barbara Hebis. 

 
2:46 
I lead the bio asset development team here in Cambridge for AstraZeneca’s biopharmaceutical development operations. 

 
2:54 
So we are the CMC arm of AstraZeneca for biologics and we develop biologics from the first manufacturing of the GLP tox slot up to the PPQ and the final pivotal trial. 

 
3:14 
And then we hand over for our commercial operations for the late stage and commercialization. 

 
3:22 
So my focus is potency assay development, but we kind of we are involved in the entire CMC process for biologics. 

 
3:35 
Good, thanks. 

 
3:36 
And my name is Laura von Schantz, the chief technology officer of Alligator Bioscience. 

 
3:41 
We are a mid-sized biotech company. 

 
3:44 
They're both monoclonal and bispecific antibodies and my background is in immuno-technology or immunology. 

 
3:52 
And well, I work with, you know, early discoverability, but I also have an overview of the clinical pathways for different molecules in my role in the management team. 

 
4:09 
But really my expertise is going to put in science to be honest, right, that's on me. 

 
4:16 
Since we're having you on the camera ratio, do you want to introduce yourself? 

 
4:24 
Sure, good morning. 

 
4:25 
I’m Horacio Nastri, I work for Inside Pharmaceuticals. 

 
4:29 
I'm in charge of protein science and technology. 

 
4:34 
So on the antibody part, we are supporting both bispecifics and monoclonal antibodies, the way from discovery up to optimization. 

 
4:46 
We interact with our CMC colleagues and so that they are in charge of all the production and so on. 

 
4:54 
But really our interfaces is very strong as we provide them the molecules and we don't want to see them back. 

 
5:02 
It was just want to see them in the clinic. 

 
5:06 
That's also our philosophy. 

 
5:08 
Yeah. 

 
5:10 
All right, thank you. 

 
5:12 
So moving forward, I will just ask anyone entering the discussion to just very short present themselves. 

 
5:21 
I was the I will run together with Barbara a short interaction the topic and then we will go into the discussion points. 

 
5:27 
I saw the answers that people had entered before the session. 

 
5:34 
I realised there's quite a lot of us interested in still in CMC and developability. 

 
5:38 
So I took a little bit of adding that as a discussion point. 

 
5:43 
All right, so a short introduction to the topic. 

 
5:46 
And I think all of you, I mean, maybe several of you are familiar with this, but a lot of things have happened the last five years when it comes to approvals for bispecifics. 

 
5:59 
But in reality we start working on bispecifics. 

 
6:01 
So the field started from 1960. 

 
6:04 
So a lot of technology development was needed before we were able to get to the place that we are today. 

 
6:13 
And this slide is just to emphasise that you know the last, currently we have 12 compounds approved by different authorities. 

 
6:25 
Half of them are in on the cancer field and we have a third of them in or more than half of them. 

 
6:33 
So half of them of blood cancers are fourth in solid tumours, which is really encouraging. 

 
6:40 
And then bispecifics are also being used for treatment of other ailments outside of oncology. 

 
6:47 
And I think Barbara will go into this a little bit more, but there are, we're still working on many different formats and we'll probably continue to do so for many years. 

 
6:59 
And we still have challenges with several other formats, but we're still getting our point. 

 
7:05 
But we are a lot of compounds on phase three, phase two and so on. 

 
7:09 
So a lot of the problems have been solved now and looking at the clinical landscape, this is taken from Beacon bispecific database. 

 
7:21 
So there is there are so many compounds behind developed at the moment in pregnant stage over 1000 molecules and almost more than 450 in clinical stage. 

 
7:34 
And so, yeah, there's a lot of common and you know people are always very innovative and a lot of compounds out there that I'm personally look forward to seeing how they pan out and how they perform in the clinic. 

 
7:52 
So I know some of you wanted to know a little bit on the applications for the bispecifics about cancer is oncology is the greatest application for bispecific still. 

 
8:05 
And we have as I said, both in development for blood cancers, but also solid tumours. 

 
8:11 
And if we'd look at the drugs that are entered the clinic this the first half of this year, solid tumours is very dominant. 

 
8:21 
So that's where we set our hopes on. 

 
8:24 
And now I live over to you Barbara. 

 
8:28 
OK. 

 
8:28 
So this is showing the slide is showing the trajectory that Laura has already pointed out. 

 
8:34 
So this is definitely a growing field, and the numbers are basically exploding. 

 
8:43 
This comes from a publication in 2021, and it already is outdated as you've seen in the previous slide. 

 
8:51 
So it divides molecules up into IgG like IgG modified and fragment based molecules and there is huge variety inside those descriptions. 

 
9:04 
So IgG like is really something that is very close to an IgG where mostly there is potential for having heterodimeric versions that are manufactured from a mix and then purified out. 

 
9:23 
So that has a significant impact obviously on yield and that's why the IgG modified versions have been developed where the heterodimer pairing is enhanced by certain mutations that enable that. 

 
9:39 
And then many of those formats have in addition to their IgG like format attached potentially drug conjugates or single chains or additional binding sites of a variety of sources. 

 
10:01 
And then there are the fragment based molecules that are as we see on the probably next slide, fragments of yeah, fragments of antibodies that are combined together to result in multi specific binders. 

 
10:17 
So this slide really illustrates the large variety and it is only a fraction of I think what is being worked on. 

 
10:26 
Probably the ones that have been given nice names have been mentioned on this slide. 

 
10:32 
So as I said, it started off with just having monoclonal antibodies produced in this two different with two different binding arms produced in the same cell line, which then results in occasional heterodimerisation. 

 
10:48 
And then just via the purification process these were purified out. 

 
10:52 
And you can imagine that in this case you get a lot of homodimers formed. 

 
10:57 
So a lot of the product is in at the end waste and only a small fraction of it is the product that you're after. 

 
11:05 
So really I think trailblazing was the trio map, which wasn't even a human or even humanised format, but it was a molecule that consists of half rat and half mouse antibody. 

 
11:20 
And the occupational heterodimers were purified out and then developed. 

 
11:25 
The development goes towards forcing heterodimerization and then potentially adding on additional binding moieties as you can see maybe coming up on the bottom left there and then at the top right you can see the large variety of where you can add single chain binders to the molecule and you end up potentially with asymmetric molecules. 

 
11:52 
So all of the ones that have added binders that are not part of the standard antibody format have to have linkers and those linkers are generally a liability sometimes already during production, but often also during stability of the products during storage and also essentially in vivo. 

 
12:16 
So there are different stages of product development where these linkers play a very large role coming up to at the end serum stability where product can actually degrade quite quickly in the recipient or even in the preclinical model. 

 
12:38 
So the main challenges really are there is a wide variety of options and somebody needs to choose the design option very early on. 

 
12:49 
And then there are purification implications where you have impact on yield if you are producing a large amount of material that you then later do not want. 

 
13:02 
And if you are adding on, you have the added downstream implications on making sure that you have a formulation and a downstream purification process that stabilises those linkers. 

 
13:20 
Thank you. 

 
13:20 
So let's come back to the design because I think that's a topic that we will dive in. 

 
13:28 
But yes, looking at the antibodies that have been approved, if we have, you know, antivirus means several different performances showing here from bites and larger molecules as you like and also appended a GG. 

 
13:43 
And so what's clear is also that, you know, people are progressing the different formats and also possibly different formats are needed for different biology anyway. 

 
13:54 
We haven't, you know, really set rules for that. 

 
14:01 
Yeah, there might be situations where one format is better than another, but let's see if we can discuss. 

 
14:08 
Yeah, there are definitely various drivers for the large variety of molecules that we're seeing. 

 
14:14 
Yes, so I also added this there, this is an old data set really, but it's still you know, in a good way visualises different reasons why would want to discontinue programmes and you know, some of the things undisclosed and some more, you know, just labour strategic decisions. 

 
14:36 
We know that there are issues with some toxicity possibility, immunogenicity and but also very importantly lack of efficacy. 

 
14:47 
So there are still challenges that need to be solved just before we jump into the discussion. 

 
14:57 
I think that there are some critical properties that need to be in place before we move forward with something to development. 

 
15:05 
And I think that in this paper by Nieto in 2020, this summaries quite well. 

 
15:12 
So you want something that has obviously the desired efficacy in the clinic, but so many more things need to be in place. 

 
15:20 
There has to be their property safety, the physical chemical properties as you were discussing Barbara and the Piki PD and ratio and the properties that need to be in place at some point. 

 
15:35 
This scalability manufacturing is not always the step forward as for monoclonals. 

 
15:41 
And as I was mentioned also immunogenicity is a handle minute. 

 
15:47 
So it would be nice to discuss and understand how people are handling that on their own. 

 
15:54 
Different companies here. 

 
15:56 
So I also added this, but I'm not going to go through it because I think that looking at that criteria, one can also set a criteria for what is inherent and for a format. 

 
16:04 
But let's not go into that. 

 
16:05 
Instead, let's look up for the discussion. 

 
16:08 
And the first topic is basically the mechanism of actions and advantages of buying multi specifics. 

 
16:16 
And before we open up for the discussion, just to be clear that when we say multi specific, we really mean both by and more than two targets at the same time. 

 
16:29 
All right. 

 
16:30 
So I just added a couple of questions from the questions that were added to them to the portal. 

 
16:41 
So let's start with discussion. 

 
16:43 
Then when it comes to the different clinical and expectations, are there any compounds that people are particularly excited to follow? 

 
16:55 
Moving on, So I can see Horatio has his webcam on. 

 
17:04 
We have none of our attendees. 

 
17:05 
I'll do one final just wave to see if you would like to join as a panellist. 

 
17:08 
Obviously you don't feel pressured to, but you can just also type stuff in the chat. 

 
17:12 
It's essentially being a panellist. 

 
17:13 
Again, just doesn't mean you can't turn your webcam on and off when you choose to. 

 
17:16 
It's just you don't have to permissions when you just join. 

 
17:18 
So I'll just do a quick wave of that, see if we get any other people pop up. 

 
17:22 
Otherwise, perhaps, yeah, Horatio would like to maybe share his thoughts on that as well. 

 
17:29 
I think that from my point of view, I'm much more interested on the discovery. 

 
17:34 
So I'm not really following the compounds on the clinic in that as we divide to conquer, this is more for the work for people doing the biology, they're really much more inclined to the indication. 

 
17:50 
So I think that from my point of view is important to understand the hypothesis and the biology that you need so that you would have the right assets to begin with. 

 
18:04 
And then you're talking about the formats. 

 
18:07 
And for me, it's not completely clear. 

 
18:12 
There may be some balances issues that two by one or two by two or one by one that may have some advantages. 

 
18:19 
But from my point of view, many of the formats were device or designed to have freedom to operate at one port in time. 

 
18:28 
And I think that as things are open it up, I think they're going to be some forms, they're going to be predominant because of what the bar I was saying regarding stability and developability and efficacy. 

 
18:44 
So I'm perhaps it's a little bit off topic, but I'm curious about what do we think that's going to be the winner? 

 
18:51 
So the predominant format, I mean, like everything is at the beginning, you have to do all the explorations, but as patterns expire, you're getting, you have multiple ways to do heterodimerization not into holes works well. 

 
19:07 
Do we need to really go more fancy than that? 

 
19:10 
Those are basically just again from the engineering and discovery part, some of the things that are important to me. 

 
19:19 
So if I can just ask you then how do you do it currently? 

 
19:23 
Do you use several formats when you're developing 1 molecule or do you try to, you know, So yeah, most of our approaches are using common like chain at this point in time. 

 
19:38 
So that it has pros and cons. 

 
19:40 
It has some a lot of easy path forward for developability and hit the administration is less of a problem. 

 
19:49 
We are exploring other ways to get basically plug and play so that we can get more diversity from other sources, but we are not. 

 
19:59 
The idea is to have that on FAB. 

 
20:02 
And we have explored some 2 by 1 and we have found that depends on the targets and the nature of your binders. 

 
20:09 
Sometimes the two by one has advantages. 

 
20:11 
And this is something that I think it has a lot of relevance and importance in there, especially for the CD3. 

 
20:18 
You may not want to have a two by two, but that's my experience. 

 
20:22 
I don't know, there was such a paper not long ago when they started this, the different balances and they still end up saying that the two by two would have been a good form for CD3. 

 
20:34 
And so. 

 
20:35 
So I don't think we know because as you say, people use kind of like the format they have and for CD3 they are they, those were toxic molecules, right? 

 
20:45 
So, yeah, that's why sometimes we try to combine the ability on the target and reducing the potency on the CD three. 

 
20:54 
That's a little bit of my philosophy. 

 
20:57 
Yeah. 

 
20:57 
I think we have learned a lot that so we don't have to go for the high affinity binders all the time. 

 
21:03 
And I can add from my perspective, we felt that the formats, even the ones that don't have been interrupted were not enough. 

 
21:12 
We work with CD44 1BB and we wanted something was 2 + 2. 

 
21:18 
And so we had to vent on format. 

 
21:25 
But I think, yeah. 

 
21:27 
And to answer your question on what is the format that's going to be the winner? 

 
21:31 
And then to me, it has to be something that has, you know, inherent good stability, inherent good illicit risk that it's easy to work also early stage in discovery. 

 
21:47 
So when you generate, when you know, combine 10 * 20 molecules, you still get a lot of molecules that are easy to work with. 

 
21:56 
So you don't lose molecules that early stage. 

 
22:01 
Well, in our experience, often it's not the basic format idea that affects stability or yield as much as we would have expected. 

 
22:13 
So we have had quite a lot of success in fine tuning the linkers for example and making small changes to the sequences that actually resulted in very significant improvements in stability. 

 
22:30 
So I at the moment, I think that a lot of formats are achievable for stability and even like in vivo stability reasons that probably weren't that achievable previously. 

 
22:48 
So our approach is to find some kind of platform ability where we have certain like modular linkers that can be switched in and out in the design phase. 

 
23:02 
So that's what I'm seeing right now. 

 
23:04 
But in terms of what's going to be the winner, I think biology plays a massive role there. 

 
23:09 
So we are seeing now formats that rely on flexibility, for example of the entire construct that then can fine tune availability of the paratope to the target under certain conditions. 

 
23:26 
So for example, binding to the tumour target then opens up the molecule to bind to the effect the target. 

 
23:36 
So I think we're going to see even more variability there. 

 
23:41 
But I also think that some basic concepts will win out at the end. 

 
23:47 
And I guess as you said, the high variability that we're seeing is caused by freedom to operate, but it will show us what succeeds. 

 
23:57 
So let me ask you something regarding what you said about the different format. 

 
24:02 
So one way to operate is that you have a large diversity of the binders, you have a simple or limited number of formats and you basically do all the combinations looking for the right geometry or combination of epitopes or the other is that you may have limited and then you try multiple structures on your antibody. 

 
24:25 
Are you saying that this is what AstraZeneca does that they explore multiple formats for during the screening? 

 
24:34 
Yes, definitely exploring multiple formats during screening, driven by biology, driven by platform knowledge. 

 
24:43 
And I have seen situations where just before candidate selection, a certain format was winning. 

 
24:50 
And then we discovered some developability risk. 

 
24:53 
And then they went back to not quite back to the drawing board, but back to an intermediate stage and optimised a different format. 

 
25:01 
And that went ahead in a very short space of time because that's kind of this modular plug and play advantage. 

 
25:10 
Yeah, I mean, I don't see the early stages of design. 

 
25:14 
So I kind of get become aware of the molecules when they approach candidate selection. 

 
25:20 
But I've seen a lot of chopping and changing right at the end as well. 

 
25:25 
So that tells me a lot of I think that it depends on the resources available that the how many extractors or how many things can you explore in a reasonable amount of time. 

 
25:38 
I would yeah, we're not able to do that than the multi-dimensional approach. 

 
25:43 
But you know, several different formats, hundreds and hundreds symbolics. 

 
25:46 
We do screen, you know, we can screen thousand but everything in the same format. 

 
25:54 
But I think that important thing then you need to have an assay that really translates into the function that you're looking for. 

 
26:05 
Otherwise you may have selected something that is not optimal. 

 
26:09 
Yeah, that's definitely a suite of assays required because you're testing for different properties of the molecule, not only for the entire mode of action, but potentially for some of the to confirm some of the design features actually have the impact that you have designed them for. 

 
26:25 
So I just want to follow up one of your comments. 

 
26:28 
You've mentioned several times the importance of the linkers. 

 
26:32 
So what are the issues you're seeing with the link of the testing? 

 
26:36 
So there is the linker that just links right, which I think has been optimised and there are some choices that we can make. 

 
26:46 
The flexibility of the linker has impact on the availability of the paratopes of whatever is attached to it. 

 
26:53 
So that's why you need a little bit of, you know, a few options there in order to choose the one that works best for the format. 

 
27:00 
And then also for the, you know, sequence where the distribution of where you put which binder. 

 
27:06 
So that's kind of the most simple linker because in most cases you have quite a lot of freedom with the sequence. 

 
27:14 
So you can eliminate sequence liabilities like deamination sites, isomerization sites and things like that. 

 
27:20 
But then you have the functional linker that may have to be cleaved by a protease or may have to allow the molecule to change confirmation upon most of the time protease cleavage or any other kind of protein interaction. 

 
27:39 
There you have very have much more limited options for the sequences because you may have an impact on function when you're eliminating liabilities, sequence liabilities. 

 
27:49 
So that's where there is a lot of linker screening and activity going on. 

 
27:58 
Sure. 

 
27:58 
Do you want to join? 

 
28:00 
Hello, Stuart. 

 
28:00 
Yes, please. 

 
28:01 
Yes, I'd like to interject with a another question here. 

 
28:04 
My name's Stuart New. 

 
28:05 
I also work at Insight Horacio's team. 

 
28:09 
I'm leading our in vivo antibody discovery efforts and my question is pertaining to like immunogenicity of different bispecific platforms. 

 
28:18 
So when you're exploring these different, you know, systems to produce bispecifics, can you comment on at what stage you would start looking at immunogenicity and what the approaches you would take either in silico or in vivo or in vitro in order to try and make some judgement as to the relative immunogenicity of these different platforms? 

 
28:40 
Yeah, I mean, certainly so there are so many different reasons why you can possibly get in university and you can't address all of them so early on. 

 
28:50 
But anything that is sequence related you can study in silly. 

 
28:54 
So that we do as early as possible. 

 
28:57 
As soon as we have a seat, we start looking at the in the risk and what we do is just rank in the beginning. 

 
29:05 
And if for some reason that clone would turn out to be, you know, superior functional compared to other clones, then we might consider moving forward and trying to reduce the message somehow. 

 
29:21 
And then before choosing ACD, we perform what is called the MAPS assays. 

 
29:28 
So we tested the proteins in dendritic cell assets as to see if they actually do lead to bind mode of HLA and I mean other reasons what you get immunogenicity or you know, aggregation oxidation and so on. 

 
29:48 
So we also look for that even the probability analysis and during CMC and just ensuring that the product is as few as possible, but you can never know until you're in the clinic. 

 
30:01 
So you're still there. 

 
30:05 
Yeah, I think our overwhelming experience is as very poor predictive power in most of what is currently available for immunogenicity testing in silico. 

 
30:19 
And also I wonder if do you have any idea about correlation between the prediction tools and whatever have seen in the clinic one way or another? 

 
30:35 
Because I've been involved many years ago with a company that was working with valuation. 

 
30:42 
So that was one of the first prediction tools. 

 
30:46 
And the question was that you never have, you never put the in the clinic the original molecule and the improved molecule. 

 
30:53 
So you don't really know how predicted are those values on. 

 
30:58 
And then I think that the in vitro is probably the best surrogate acid you have. 

 
31:04 
But as you said they are also with the caveat there are also some both system are over predictive unfortunately. 

 
31:13 
But what we have done is we have, you know, going back, if we find out that a molecule is immunogenic, not necessarily our molecules, but just in understanding that molecules when they try to see if it would have been picked up by the in silico assay. 

 
31:28 
And actually we do see that, yeah, but we rarely see the opposite. 

 
31:34 
So what we don't see is molecules that are not immunogenic, sorry in the clinic to come out clean out of an immunogenicity screen. 

 
31:46 
We I don't think we've ever seen. 

 
31:48 
So not it's always in relation to something else. 

 
31:53 
So those facts they cannot say if it's generic or not, but you can say among 1/3 molecule you have this one have a higher risk than this one. 

 
32:01 
Yeah, but it's often also dependent on the mode of action. 

 
32:04 
So obviously if you have molecules that engage the immune system, you often see immunogenicity where sometimes you wouldn't see it in a molecule that doesn't do that, even though it has the same sequence in an exposed position. 

 
32:19 
Definitely. 

 
32:20 
Yeah. 

 
32:20 
Yeah. 

 
32:21 
But the dendritic cell assays are quite good, at least in predicting the processing, because most of the prediction sequence predictions are independent of the neighbouring sequence. 

 
32:34 
And yeah, that's very good. 

 
32:40 
So what about PK? 

 
32:42 
So I mean, you could say that several or some candidates have been dropped clinical development because of PK and there is any way that we can more or less predict or have some experience about PK, especially with bispecific monoclonals may be a little bit easier. 

 
33:03 
But so we regulate tests done in mice to be honest, early on before we choose the CD, we try to test the top candidates to see if there's any for any reason, something unpredictable that is often or sometimes difficult if you have a target that will affect the PK. 

 
33:27 
But where it's possible, we try to test the PK both 1st and mice and then obviously monkeys. 

 
33:37 
Are you using just well type mice or FCR and humanised mice? 

 
33:44 
So all of them are valuable just to choose one clone over another one. 

 
33:53 
One type mice work well. 

 
33:57 
But if you want to, you know, study a particular feature of your format compared to another monoclonal, then it's better to use the FCR. 

 
34:10 
Yeah. 

 
34:12 
When you're doing these in vivo PK studies, do you find that the bispecific platform universally affects the PK properties of different variable regions or do you find that it that really boils down to a case to case situation? 

 
34:26 
And so no, I mean, you look at the literature, we and we haven't tested that many other formats. 

 
34:31 
We have our format, we'll be and that, I mean, behaves normally like a monoclonal antibiotic. 

 
34:37 
And so we don't see a difference in PK because of the format. 

 
34:40 
But if you look at the literature, you can see that there are formats they are less stable. 

 
34:45 
If they are less stable in general, then they also normally have less, you know, a faster PK or faster halfway. 

 
34:54 
Yeah. 

 
34:55 
And obviously when you have multi specifics, bispecifics I think are maybe slightly more predictable and similar to maps as long as they have a different a similar format. 

 
35:08 
But with the multi specifics you always add potential for target mediated disposition. 

 
35:14 
So your PK predictions generally are wrong, but the question is, are they indicative or not what to expect in human PK? 

 
35:25 
That's generally the question. 

 
35:27 
And we have with these more out there formats seen very unexpected mouse PKS and I don't think we have enough human data to know what the predictive power of those observations was so far. 

 
35:46 
Yeah. 

 
35:47 
So I mean, I would say you can use the mind small mouth models to understand the stability and the clone differences, but you can't really understand the biological TK differences either in mice or in Sina. 

 
36:02 
You can prove yourself with science studies as well. 

 
36:05 
Yeah. 

 
36:06 
Often you don't have the same distribution of the, for example, use the TA that's not the same way. 

 
36:12 
Exactly. 

 
36:16 
Yeah, but that's tricky. 

 
36:19 
So how do you guys do it? 

 
36:23 
Yeah. 

 
36:24 
So we are also doing mice and humanised mice and Sino. 

 
36:28 
But sometimes we are seeing abnormal behaviour and that is difficult to interpret. 

 
36:37 
We also have been seeing a lot of ADA being produced both by Mass and Sino. 

 
36:46 
And then it's something that it basically obscure your understanding of the results one way or another. 

 
36:52 
So I mean those are experiments you need to do, but actually you need to proceed regardless of what they say. 

 
37:01 
That is a right. 

 
37:07 
So you don't get it either. 

 
37:12 
I'm not sure if they are skid mice. 

 
37:15 
I don't really know. 

 
37:16 
Do you know us too? 

 
37:18 
I don't believe that they are. 

 
37:20 
But as Horacio mentioned, it's a little bit downstream of our typical involvement. 

 
37:26 
So I think there was a version that is the human eyes we use. 

 
37:32 
I don't think that there is skid. 

 
37:33 
I think that there are, that they have their, there's just no key in mice and they have the regular immune system. 

 
37:40 
But yeah, I mean AA wrong about that. 

 
37:44 
ADAS play a big role in your efficacy model obviously. 

 
37:48 
And often the ADA assessment informs more your interpretation of your efficacy model than it does inform any future, any clinical ADA problems. 

 
38:01 
Yeah, sort of in line with the question of the FCRN mice and PK studies. 

 
38:09 
Do either of you see a lot of utility in doing more like SPR analysis of FCRN binding for your molecules before you get to an in vivo situation or do you just go straight to the in vivo model? 

 
38:25 
I have a very strong, I have a very strong opinion on this topic. 

 
38:29 
Oh, really? 

 
38:31 
OK, yeah, we always do on binding studies before. 

 
38:38 
Yeah. 

 
38:40 
So in my opinion, my human and different pH of course. 

 
38:46 
So you both feel strongly about the utility of those assays in the development process? 

 
38:53 
Well, so from our MAP experience, I have a very strong opinion about that because we know exactly what FCRN binding is dependent on. 

 
39:05 
And that's the status of the methionines and the oxidation of the methionines directly correlates with FCRN binding and it directly correlates with FCRN only PK. 

 
39:18 
But obviously in the clinic, the PK, the FCRN only portion of the PK is almost, you know, has very much less relevance than it has in your models. 

 
39:33 
So I never understand why we in BIOS they have to do so many FCRN binding studies. 

 
39:42 
We already know the methionine oxidation under stress. 

 
39:47 
So that's the one thing. 

 
39:50 
And then in general, as I've just said, the more specificity, the more target binding opportunity you add to a molecule, the less predictive is your FCRN portion of the PK. 

 
40:08 
So you're doing them, but you think it's highly predictable and the translation to the actual inpatient PK is limited? 

 
40:18 
Well, yes, as long as you have FC containing formats. 

 
40:22 
We do our binding and we try and avoid the forced degradation sample explosion, but I don't think the value of it is that high. 

 
40:38 
But you know, it's one of those things that almost ends up as a tick box exercise in your IMD. 

 
40:43 
Sure. 

 
40:43 
And I think that is what we are trying to avoid in general. 

 
40:50 
There are many things we could do, but the question is that makes any sense to do it or not. 

 
40:56 
So, yeah, I mean even in mouse PK models in the absence of target with human FCRN, we see with those different formats we see unexpected PK where you have no target media disposition component involved. 

 
41:18 
So yeah, that tells you have polyspecificity on off target binding or you don't think that you have that. 

 
41:28 
Well, it's difficult to tell because I don't think we do mouse tissue crosser activity. 

 
41:35 
All right, Yeah. 

 
41:37 
So do you ever see a correlation between the Pi and the PK? 

 
41:45 
I honestly don't know. 

 
41:48 
I don't think I have. 

 
41:50 
I don't think I have scrutinised that data and correlated it. 

 
42:00 
There's also the assumption that glycosylation has an impact on PK. 

 
42:06 
And again, there's very little evidence, I think also in the literature. 

 
42:11 
But you always get this. 

 
42:14 
Yeah. 

 
42:14 
The balance, even if the A is interested in that. 

 
42:17 
Exactly. 

 
42:18 
Yeah. 

 
42:20 
Well, I mean, you have mutants, though, in theory, enhance PK, don't you? 

 
42:27 
Yeah, we have, Well, we have a mutant that enhances the binding to FCRN, for example, that we generally use for molecules that we want to stick around for longer. 

 
42:40 
Has some drawbacks when you combine it, for example with effective function null mutations, some drawbacks on the stability of the molecule. 

 
42:49 
Then you've got ADC's where you have got conjugations conjugate if impact on conjugation, impact from conjugation on stability that then conflates and obviously you're losing one of the methionine. 

 
43:04 
So you actually don't get the predictive on one hand, you don't get the predictive power of the methionine oxidation. 

 
43:11 
On the other hand, you know that you've got your mutation and you've got, you know, you've got your enhanced binding. 

 
43:18 
And then you can in some cases you can actually surprisingly, I think counteract PK impact by adding the enhancing mutation and bring it back to more MAB like PK. 

 
43:35 
I see, at least in models. 

 
43:38 
Yeah, OK. 

 
43:41 
And you don't typically get immunogenicity again, those mutations, because that's always the risk. 

 
43:47 
Yeah, because we have validated mutations, clinically validated mutations. 

 
43:54 
We use those. 

 
43:56 
I see. 

 
44:02 
And what people always forget, I think at least that's my impression, is that if we have an FC in our molecule, it's always the identical life. 

 
44:09 
All our molecules have the identical FC portion apart from those, you know, deliberate mutations. 

 
44:18 
So we have loads of data. 

 
44:20 
So I don't know why we keep having to assess it for each molecule since we haven't made a change in that area of the molecule. 

 
44:30 
So you may need to ask a regulator why they are asking. 

 
44:33 
Well, that's the thing. 

 
44:34 
We have done that actually. 

 
44:35 
I'm currently on a panel on a cross company panel writing a white paper on that aspect and also on some other aspects that the regulators keep. 

 
44:47 
I mean under a different effort that the regulators keep asking for no scientifically founded risk. 

 
44:55 
So what are the other things that you feel the regulator ask for? 

 
44:58 
Well, this is a totally multi specific unrelated issue, which is where you have your target is a soluble molecule. 

 
45:09 
And so therefore why do you have to have a cell based assay to measure the potency of that when it doesn't interact with cells at all. 

 
45:16 
But it prevents the binding of your soluble molecule to its receptor, which you can very nicely measure in a biochemical assay, in a competition assay. 

 
45:26 
But it keeps ending up with this question when you're coming close to pivotal trial, you have to have a cell based assay. 

 
45:33 
And we say, no, you don't, but you know, sometimes you lose. 

 
45:38 
And what do you do then? 

 
45:40 
What do we do about it? 

 
45:42 
Transfect cells with the target, even though. 

 
45:44 
Well, no, yeah, the target is soluble. 

 
45:47 
So you just do. 

 
45:49 
Yeah, you just do a cell based assay where you measure the impact of the what you've already demonstrated my soluble mediator, when bound to my drug, no longer binds to its receptor. 

 
45:59 
So therefore there's no signalling. 

 
46:00 
So you show that there is no signalling, but you've already shown it doesn't bind. 

 
46:03 
So where does the signalling? 

 
46:04 
Where's the signalling supposed to come from? 

 
46:06 
It's because, you know, the sinking comes from an agon perspective. 

 
46:10 
Yeah, you can't. 

 
46:11 
If your molecule still binds, you could still see no signalling. 

 
46:15 
So your Organism is not confirmed by binding, but your antagonism is by it's obviously confirmed by no binding. 

 
46:23 
I mean, we're not talking here about allosteric. 

 
46:25 
We're talking here about blocking your interaction directly between your ligand and your receptor. 

 
46:34 
All right. 

 
46:34 
Are there any other topics you would like to discuss here? 

 
46:40 
I know some people were interested in like future directions, so we just dwell on that a little bit. 

 
46:49 
What is coming next, you think? 

 
46:53 
There's all sorts of wild and wonderful stuff in the pipeline. 

 
46:56 
I'm really excited about Corpus. 

 
46:58 
Can't disclose, but the, I think the Pandora's box has been open with the bispecifics and then all these different formats are coming out now where you get this added functionality, recruitment, targeted recruitment, localised recruitment of immune effector function. 

 
47:20 
I mean, it's, yeah, it's wide open. 

 
47:24 
As long as we get, you know, as long as we get the molecules to be stable enough and to be to have a significant yield, I think there'll be some exciting molecules coming up. 

 
47:38 
And I think it's the end of the CAR T cell. 

 
47:40 
But that's just my opinion really. 

 
47:45 
I used to think that. 

 
47:45 
Now I'm really convinced that they are there to stay. 

 
47:50 
Really. 

 
47:51 
Yeah, they're far too expensive. 

 
47:53 
Yeah, they are very expensive. 

 
47:54 
They are right now I think yeah, I think that the really improve delivery is a big problem. 

 
48:07 
I think that we are using infusions IP working and but I think that on the long term really getting better way for delivery either sub Q or other, it's going to be necessary because if not there's going to be always a limitation about the willingness of people to take it. 

 
48:31 
I mean, if you have to go and get a half a litre or two litres infusion once a month, that's not really a very good or nice quality of life. 

 
48:40 
Yeah, I was already sorry. 

 
48:43 
Carry on Laura. 

 
48:44 
I was going to say that to get the thick as if we're not currently getting with all the bispecific it's a bio distribution problem. 

 
48:52 
So we're not getting the drugs in the tumour in the amounts we want to. 

 
48:59 
Yeah. 

 
49:00 
And it could be that we need to do it a tumoral administration, but I know it's a high resistance to that. 

 
49:08 
But there are also people, you know, evaluate other routes, you know, vaccination of the liver systems, virus, the liver system and so on. 

 
49:20 
I think they're many years behind that. 

 
49:23 
Yeah, they will also probably get there at some point. 

 
49:27 
I think combination therapy with maybe chemo, adjuvant therapy after surgery, things where you have a disrupted tumour microenvironment, you have a better chance to get in with a drug. 

 
49:44 
So there are these studies that are also are happening. 

 
49:50 
Yeah. 

 
49:56 
I think in from the technologies we're still seeing, you know, more advanced specifics, stride specifics, sector specifics. 

 
50:06 
I think they have a lot of challenges ahead with I mean we still haven't solved all the difficulties with the dosing of the bispecifics for example. 

 
50:16 
Then I can't imagine how difficult it must be to dose that right specific and how to balance affinity with it there all the, those parameters and target mediated disposition. 

 
50:27 
Because if you have a bispecific, you may have one target that is abandoned as a the thing as a single express target in normal tissue. 

 
50:38 
And then your other target may be more restricted to the tumour. 

 
50:42 
But you have a massive sync for your molecule where it doesn't do anything. 

 
50:47 
And then that's, we're currently working on a molecule where the entire organisation keeps asking us why do you have two molecules? 

 
50:54 
Why do you have two molecules? 

 
50:55 
Can't we do it in one? 

 
50:56 
Yeah, we can technically do it in one, but we think it may be an advantage for those things. 

 
51:02 
So we'll see. 

 
51:05 
No, I think that I mean for a long time we said in the field that you have to have something that needs to be very specific to be worked to it has to be obligated. 

 
51:15 
But I think it's the same rule that applies for Tri specifics and types of specifics. 

 
51:19 
I think there are some design ideas coming I guess into the clinic maybe that are not you know that are not circumventing the bispecific format, but they are maybe adding it on as an effector. 

 
51:41 
So where you have, you know, we have separate molecules and they actually come together in the tumour microenvironment and then they have their bispecific function or they have a bispecific and recruiting effector function that is activated in the tumour microenvironment. 

 
52:01 
And it has to be the final format to be the final format. 

 
52:06 
Yes. 

 
52:06 
OK. 

 
52:06 
So then it's still but not your actual drug. 

 
52:10 
Yeah. 

 
52:10 
No, exactly. 

 
52:12 
Yeah. 

 
52:16 
Interesting. 

 
52:16 
So is that something you're developing? 

 
52:19 
Well, there's definitely something that is on the drawing board. 

 
52:22 
Yeah. 

 
52:22 
Yeah. 

 
52:23 
No, I thought it was something. 

 
52:24 
No, I don't need for you to share something with that. 

 
52:29 
All right. 

 
52:32 
Any other questions? 

 
52:37 
No. 

 
52:39 
Declan, are you there? 

 
52:42 
We wrap up. 

 
52:43 
Yeah, I'm happy for you guys to wrap up, I guess. 

 
52:45 
Yeah. 

 
52:45 
Do you guys have any closing remarks or anything like that? 

 
52:47 
Otherwise, I can, you know, just close up. 

 
52:50 
I think there are no questions in the chat currently. 

 
52:52 
So Horatio Stuart, if you're all good, I'd like to say a big thank you to Laura and Barbara for, you know, participating as leaders for this. 

 
52:58 
It's much appreciated and also much appreciated, Horatio Stewart, for joining a webcam as well. 

 
53:03 
So it's always great when people get more sort of involved that way. 

 
53:06 
Otherwise, you know, thank you to all our audience as well, participating that way. 

 
53:10 
It's great to sort of get you all involved. 

 
53:12 
And yeah, that's like to welcome you to our next session so well in the future. 

 
53:17 
So thank you, everyone and a very great rest of your day. 

 
53:20 
And we'll see you at the next future sessions.